Corrections or additions?
Meet the Computer with the Working Heart
This article by Christopher Mario was published
in U.S. 1 Newspaper on May 13, 1998. All rights reserved.
When F. Scott Fitzgerald famously declared that there
are no second acts in American lives, he wasn’t counting on guys like
Bill Scott. After leaving his job as head of worldwide research for
Bristol Myers-Squibb in 1996, Scott, now 58, could have done a number
He could have returned to academe, having been a researcher at the
prestigious medical institution Rockefeller University for 16 years
before jumping to industry in 1983. He could have joined a bunch of
boards of directors. He could have become a consultant. Or he could
have landed on some tropical island somewhere under his golden
of heftily appreciated Bristol Myers stock options, there to live
out his days in splendid and well-earned repose.
Instead he signed on as CEO of a start-up biotech company with a
so advanced it sounds like science fiction.
Scott’s company, Physiome Sciences, has a $2 million supercomputer
purring away at its College Road offices. Performing dozens of
equations simultaneously, the computer holds within its electronic
brain a fully functional, three-dimensional, interactive model of
a working heart. The result of 30 years of research and made possible
only recently by significant advances in computing power, Physiome’s
"virtual heart" is expanding the horizons of understanding
of the organ whose various malfunctions are together the leading cause
of death in the United States (http://www.physiome.com).
This heart is not just some high-tech computer game. And it’s not
just an animated picture of a beating organ, although it does offer
a graphical interface. Rather, Physiome’s virtual heart is an
detailed math-based model of how the heart actually works. Not just
that it beats, but why and how it beats.
Based on what until just a few years ago had been the entirely
work of an Oxford physiologist named Denis Noble, the virtual heart
provides a quantitative, comprehensive, verifiable, and usefully
model of the heart based on the properties of the heart’s individual
cells and the biochemical functions they perform. Translation: the
virtual heart breaks down the millions and millions of biological
processes going on inside the heart, describes them as mathematical
equations, and then assembles them and all their actions and
to create an electronic heart that "works" just like a real
That alone would be impressive, a technological feat that even five
years ago was pure science-fiction, the medical equivalent of cold
fusion. But Physiome’s virtual heart does much more than merely mimic
the function of the heart. Because the known mechanics of the heart’s
most basic cellular functions underlie this virtual heart, it’s
to reprogram the heart by changing the equations that describe those
Which means that Physiome researchers can program their virtual heart
to have a heart attack, or develop congestive heart failure, or go
into arrhythmia. Stuff that happens to real hearts, inside real
And that’s what makes Physiome interesting. If you can create a model
that demonstrates what goes wrong in a heart based on the actual
and biochemical processes responsible, you’re that much closer to
figuring out how to fix it.
You have probably never given much thought to how and
why your heart works, but Bill Scott will tell you anyway:
"The heart is an electrical organ," Scott explains. "What
drives the heart to contract and relax is the movement of calcium
and other ions in and out of cells. As a result of that ion movement,
you get voltage movements across the surface of the heart. That’s
what an electrocardiogram (ECG) measures. When you put electrodes
on a person’s chest, what you see is a complicated series of peaks
and valleys, which represent voltage moving across the heart."
Physiome’s virtual heart is built on mathematical descriptions of
the major gene products of the cells of the heart. Most of these gene
products regulate ion channels — routes across which calcium,
sodium, potassium, and other ions move in and out of cells, and in
so doing create the electrical charges that power the heart. You need
a supercomputer to do this — Physiome has a Silicon Graphics 32
parallel processor — because we’re talking about a lot of gene
products, a lot of ions, and literally millions of equations to
them all — what computer types call "embarrassingly parallel
"These descriptions are based on experimental data, and are
in the sense that taken together, they behave correctly," Scott
says of the mathematical descriptions of the ion channels that make
the virtual heart work. These descriptions, all of which are based
on actual physical data collected in experiments on real hearts by
literally thousands of scientists over many years, even enable
virtual heart to generate its own ECG.
Much is already known about these ion channels, which
have been a major area of cardiovascular research for some time. That
research has led to a discovery you have probably heard about if you
or someone you know has heart disease: the calcium channel blocker.
A class of drugs like Pfizer’s Procardia XL, calcium channel blockers
inhibit the transfer of calcium ions across cell membranes in the
heart and other muscles, causing the heart to relax and the arteries
to dilate, thereby reducing angina and high blood pressure.
Research on ion channels has also shown that in a failing heart, the
gene products that regulate the movement of ions in and out of cells
change in four major ways. Physiome used that data to reprogram its
virtual heart to behave like the heart of someone with congestive
heart failure, a common condition in which the heart gradually loses
its ability to pump enough blood to satisfy the body’s needs.
"We went into our model and introduced those changes, and sure
enough, the model then behaved like a failing heart," Scott
"In the normal heart, when it contracts you see smooth movement
of voltage across the heart. In the congestive heart, you see major
arrythmias, which appear as spiraling waves of voltage changes, rather
than smooth movements of voltage, on an ECG. Basically, a heart in
congestive heart failure gets twitchy, and though you probably think
heart attacks kill the most people, 50 percent in fact die from this
kind of electrical instability."
Physiome’s ability to change their virtual heart from healthy to
proves the model works, Scott says. If you know that four cellular
changes occur in congestive heart failure in a real heart, and then
you program your virtual heart to include these four changes, your
virtual heart should exhibit exactly the same symptoms as the real
heart. And that’s just how it worked.
Even the Food and Drug Administration (FDA), which regulates and
all prescription drugs in the United States, is convinced. Data
by Physiome’s virtual heart recently saved a new calcium channel
submitted for review to the FDA by Swiss drug giant Roche from being
"Just as they were finishing clinical trials" — testing
in humans — "they ran into a problem," Scott says of
the new Roche drug. "A small percentage of patients on the drug
developed abnormal ECGs. In just three weeks we demonstrated that
the changes were benign, and presented our findings to the FDA
committee considering the drug. Our findings turned around the
Most of the questions were about the ECGs, and we were able to remove
that as a block to approval."
Which points up the most exciting potential use of the virtual heart:
to identify new potential drugs, and then to test them.
Most drugs today are developed in a process called rational drug
Rational drug design uses a wide variety of technologies to elucidate
chemical functions that cause disease, and then seeks to identify
within those chemical functions likely targets for drug intervention.
Once the target is identified, drug researchers then search for a
chemical compound that will either inhibit or encourage the particular
biochemical activity identified as the target. Because Physiome’s
virtual heart is based on the basic chemical functions of the heart’s
cells, Scott believes it will enable researchers to tap a heretofore
unreachable cornucopia of new drug targets for heart disease.
"Our model has the descriptions of all the biochemical and
aspects of the gene products controlling heart function," Scott
says. "Which means we can inhibit or stimulate each of those gene
products to look for changes that will be efficacious in treating
At the same time, the virtual heart also provides a novel way to test
the efficacy of new drugs — a way that’s faster, cheaper, and
less ethically troubling than traditional animal testing. Just program
in the chemical changes your drug causes, and let the computer tell
you what happens.
"We can look at drug effects. We can look at dose response,"
Scott says. "This technology dramatically increases the speed
with which researchers can do their experiments, and enables them
to learn new things that they could not learn on an actual heart.
And it cuts down on the need to use experimental animals, which is
not only a cost issue but an ethical and emotional issue as well."
A third major benefit of the virtual heart is in the area of
defibrillators, better known as pacemakers.
"People with defibrillators have a pretty horrible life,"
Scott says. "These are very sophisticated devices that have to
sense when the heart is beating abnormally, and then shock it back
into a normal rhythm. The problem is, they often go off spontaneously
and put the person into arrhythmia. So people live in fear of their
pacemakers going off. A lot end up in therapy."
Scott believes his virtual heart will eventually enable technicians
to pinpoint exactly where a pacemaker should be installed, and exactly
what it should do. Scott calls the virtual heart a "rational
tool" for implantable defibrillators that will enable the optimal
placement of electrodes and the development of just the right timing
and type of shocks — the "optimal shockwave form protocol"
— thus making pacemakers less likely to malfunction.
The advent of relatively cheap supercomputing power
has made Physiome’s virtual heart — and ideas like using it to
improve the lives of people with pacemakers — possible. But the
real genesis of the virtual heart occurred nearly 30 years ago in
England, long before supercomputers were even imaginable.
That was when Denis Noble began his career as a physiologist, and
devoted himself to figuring out exactly how the heart generates its
electrical charges. A pioneer in the field of integrative physiology
— taking all the disparate bits of data about cells’ functions
and integrating them into a model of how a cell or tissue or organ
works — Noble had been building single-cell models of cardiac
tissues for nearly three decades when he and another Oxford Ph.D.
named Jeremy Levin decided to explore how advances in computer
might further Noble’s work.
Levin, who has run a number of biotech companies in his career,
together Noble and a Johns Hopkins biomedical engineering and computer
science professor named Raimond Winslow. Noble and Winslow joined
forces, and with the later addition of a team of bioengineers from
New Zealand, created the first 3-D model of a human heart based on
Noble’s pioneering research into the functions of cardiac cells.
That was in 1993. By 1996, Levin and the scientists were ready to
form a company to commercialize their invention. They raised a $2.5
million in seed funding, and a year later, they hired Bill Scott.
Scott, the oldest child of six, grew up in Illinois, where his father
ran a hardware store. He majored in chemistry at the University of
Illinois, Class of 1962, and has a PhD in biochemistry from CalTech
with postdoc studies at Rockefeller University, where he taught for
16 years. He and his wife Lonna, a freelance medical illustrator,
have a grown daughter who is an attorney.
When he was the head of research for Bristol-Myers Squibb, he had
served on the board of a company called Cadus; its CEO was Jeremy
Levin. When Levin asked Scott to join in his new venture, Scott needed
"I got to know Jeremy well at Cadus," a publicly-traded
that has a yeast-based drug receptor identification technology useful
in rational drug design, "and it was clear to me that he had a
very unique technology. The most unique thing I’d seen anywhere."
Scott also saw in Physiome Sciences an opportunity to make a
"I wanted to go back and do something hands-on, rather than just
be a high-level bureaucrat, which was basically my job at B-MS,"
Since Scott joined, the company has raised $10 million to fund its
operations, chiefly from Oxford Bioscience Partners in Westport,
but also from SR1, the investment arm of Smith Kline Beecham, based
in Radnor, Pennsylvania. This is enough to last two and a half years
without additional funding (which, like all start-ups, it continues
The company has a functioning single-cell model that can run on
NT, it has the heart model, and is beginning work on a virtual kidney.
Physiome has 10 employees in Princeton; Scott expects to have about
25 by year’s end, and spends much of his time these days interviewing
candidates. Like most employers these days, he’s having a hard time
finding good software engineers, but the people who understand the
theory — the computational biologists — are contacting him,
"It’s the first commercial application in their area," Scott
Running his small start-up is fun, Scott says. Without a huge
it’s easy to get things done. But no infrastructure also has its
Last year the company was dealt a big blow by the state of New Jersey
when the state reneged on space it had promised to Physiome at the
New Jersey Technology Center in North Brunswick. Originally slated
as a high-tech incubator, the Center was instead leased to Merck (U.S.
1, February 25, 1998).
"That set us back two or three months," Scott reports. "We
had a signed deal, and there were a lot of incentives to go there,
like money for build-out, and suddenly it was gone. They called and
The space problem solved with more expensive digs in 8,000 square
feet at 307 College Road East, Scott and his team are now focused
on seeking partnerships with drug and device research companies.
will not lease or sell its software, but rather will enter into
agreements with research teams at client companies, playing a role
in their work using Physiome’s virtual organs and single-cell models.
As for an IPO, Scott says that’s very premature. "We’re currently
very focused on three areas," Scott says. "One is hiring
Two is working on developing our corporate partnerships. And three
is spending a lot of time out presenting the company. A major issue
for any small company is becoming known, and luckily for us that’s
very easy, because what we’re doing is so novel."
Physiome has gotten a ton of press in the past year: in addition to
the many medical and technology conferences at which he speaks, Scott
has also appeared on the BBC, German television, and CNN Financial.
Whether scientists or laypeople, everybody is amazed by the computer
with a working heart.
08540. Bill Scott, CEO. 609-987-1199; fax, 609-987-9393.
The intravenous solution you receive in a hospital
didn’t start out as a liquid. A nurse took the powdered form of the
drug and liquefied it, using a needle to inject it into the IV bag.
If that sounds like a situation liable to infection, you’re right.
Raymond J. Scheire has a product that improves the safety and
of this crucial process. His Bio-Set line of drug delivery systems
enables pharmaceutical companies to differentiate their products to
more effectively compete with generics.
Scheire has opened a sales office at 5 Independence Way for Biodome,
which is based in Issoire, France. The firm’s CEO is Jacques Gardette,
and Scheire is the sales and marketing director of Biodome America
"Our product is a device to reconstitute a powdered or lyophilized
(freeze-dried) drug," says Scheire. "It eliminates a
different components, it comes in one piece, and it comes with the
drug. Before you inject a drug you can reconstitute it without using
needles." Other advantages are prevention of needle stick injuries
and reduced potential for contamination.
Scheire believes his needleless Bio-Set products are more competitive
than those made in the United States. One of the products replaces
vial and syringe systems, and other reconstitutes powdered drugs into
Bio-Set costs from 50 cents to $1.50 per unit, depending on volume,
and it is targeted to expensive drugs — such as oncology drugs,
certain vaccines, and drugs for multiple sclerosis, cardiovascular
disease, and hemophilia — which can cost from $100 to $500 per
"Our clients are looking to add value with a delivery system that
these drugs deserve," says Scheire. "With a needle and a
it is difficult to get a precise dosage. With two or three milligrams
worth $500 you don’t want to waste anything. It is awkward even for
an experienced nurse, but with our system you can be precise."
"Our device acts as a closure system to crimp the stopper on the
glass vial to have a hermetic, sterile system," he says. "The
key advantage of our system is that it can fit a traditional vial
and rubber stopper, so all we are replacing is the crimp. You need
a capping machine to push the Bio-Set on the vial. All our stability
studies are to prove it has the same closure integrity as the aluminum
The increasing number of hospital infections helped to drive the move
toward needleless systems. In 1992 the American Society of Health
Care Pharmacies issued a mandate to use needleless systems as much
as possible, and hospital pharmacies started demanding them. "We
are not pushing a product, we are meeting market demand," says
Scheire. "If a drug company offers a delivery system and is
to absorb the cost, it can go to hospital pharmacies and say, `We
are offering a delivery system that will make it easy to reconstitute
the drug,’ and that will increase market share, which offsets the
cost of the machine."
The products are being launched in 26 countries, and in the first
quarter of next year one is scheduled to be used by a New Jersey-based
pharmaceutical firm (unnamed, as yet) for the reconstitution of an
anti-infective drug administered with an IV bag.
Biodome has 70 percent of the hemodialysis market in France, and it
is distributing the products of a Japanese firm, including dialyzers,
hemostatic bandages, disposable blood line sets, catheters,
fistula sets, and single patient dialysis units. It also distributes
polio vaccine systems in Europe under contract with the World Health
Its clients include Hoechst Marion Roussel, Merck Sharp & Dohme,
Merieux, and Rhone Poulenc Rorer, Smith Kline Beecham, and
Squibb, and Wyeth Ayerst, among others.
Born in Belgium, the 40-year-old Scheire migrated to Sydney,
and then followed his future wife to the United States after she began
working in New York. He earned his BA and MBA in Sydney, has a
degree in food science from Rutgers’ Cook College, and worked for
a small company in the packaging industry before joining Biodome last
year, "To function in my position, it helps to have had a varied
background," says Scheire.
Princeton 08540. Raymond J. Scheire, sales and marketing director.
609-514-5170; fax, 609-514-5171. Home page:
Corrections or additions?
This page is published by PrincetonInfo.com
— the web site for U.S. 1 Newspaper in Princeton, New Jersey.